Using GrADS

  GrADS can be used for simple manipulation and display of mean and standard deviation data from the MCD and for more complicated calculations of derived quantities, e.g. vorticity. GrADS can also be useful for quickly translating portions of the DRS format files into other formats.

N.B. Some GrADS routines, in particular hdivg and hcurl which compute horizontal derivatives, assume that the data relates to the Earth; if they are used for calculating derived quantities from the database, results from them should always be multiplied by a factor of approximately 1.88 to account for the relative radii of the Earth and Mars. The dates used for the GrADS scripts are in Earth format, and are arbitrary, but the universal time of day is correct with one ``hour'' corresponding to 1/24 of a Martian Solar day (a Martian Solar day is 88775.2 seconds, so here 1 ``hour'' is actually about 3699 seconds). In the database 00:00 is midnight and 12:00 is noon at 0$^\circ$ longitude in True Solar Time.

A simple interactive GrADS session to plot the seasonal average surface temperature for season 4 from the Viking scenario simulation may run like this:

ga> open viks05me
Scanning description file: viks05me
Data file viks05me.dat is open as file 1
LON set to 0 360
LAT set to -88.125 88.125
LEV set to 0.9995 0.9995
Time values set: 2001:1:1:0 2001:1:1:0
ga> set mpdraw off
ga> display ave(tsurf,t=1,t=12)
Averaging. dim = 3, start = 1, end = 12
Contouring: 150 to 230 interval 10
ga>

To produce hard copy, type enable print filename.gx, then display the data which you require and type print when you have what you want. A clear command will start a new page and disable print or quit at the end of the session closes the filename.gx file. The .gx file can be viewed again on the screen with the gxtran utility or converted to PostScript for printing with gxps or gxeps. These utilities, and instructions for using them, are provided with the GrADS distribution (whether you get gxps, gxeps or other programs with similar names depends on exactly which version of GrADS you obtain).

For more complicated output, or for repeated applications which can be run in batch mode, it is more convenient to write a GrADS script. Some are provided in the mcd/grads subdirectory; it might be necessary to change the path to the data files to where the mcd directory has been installed on your machine if these scripts are to be run from another directory. The final part of the filename determines the dust scenario and season according to the code described previously in this document. These can be changed, and the titles modified accordingly, by editing the scripts. If different quantities are required from those which they plot, it should be straightforward to write new scripts using the ones supplied as templates. Most of the scripts are intended to produce output in portrait orientation, so either run GrADS and answer no to the initial landscape mode question, or specify portrait mode from the command line, grads -p. It is also useful to run the GrADS command set mpdraw off, which prevents GrADS from drawing outlines of countries on the Earth, although the scripts here will all do this explicitly when necessary; this could be achieved with a command aliased to grads -pc "set mpdraw off" which would run GrADS in portrait mode with no map outlines.

The scripts which are provided in the mcd/grads subdirectory are briefly described below. The scripts have been commented and they should provide a useful starting point.

• anim_tsurf.gs A simple animation of the diurnal cycle of surface temperature means during one season. Useful for a quick check that the data have been read correctly; a warm region (local afternoon) should be seen to propagate from East to West across the map.
• map_ps.gs Plots maps of the diurnal mean surface pressure and its (non-diurnal) standard deviation. GrADS should be run in portrait mode for this script.
• map_windt.gs Plots maps of the diurnal mean near-surface wind and surface temperature. GrADS should be run in portrait mode for this script.
• profile.gs This script gives the mean, and confidence limits at a chosen number of standard deviations, for surface variables and also plots the vertical profile of three dimensional variables from the Mars Climate Database. It is important to note that the standard deviation merely represents the variability between profiles at that location and the same local True Solar Time of day during that season. It does not include any diurnal variability (this is represented in the twelve means contained in the mean data file for different times of day) nor any small-scale variability nor any information about correlation and scales of variability within or between individual profiles (the last two points are addressed by the main database software).

  The longitude, latitude and local True Solar Time are supplied as arguments. The longitude should be given in degrees East and the time in hours (00..24). This script must be edited to select a dust scenario and season and to change the title accordingly.

  A metafile, named using the plot parameters is written, profile_lon_lat_hr.gx (where lon is the longitude, lat the latitude and hr the local time selected), which can then be viewed with gxtran or converted to postscript for printing with gxps or gxeps.

  GrADS should be run in portrait mode.

  Sample output from the profile.gs script is shown in Figure 1 for summer afternoon profiles near the Viking Lander 1 site and for contrast Figure 2 shows night-time profiles in a winter polar region at the same time of year.

  
  

    

  Figure 1: Sample output from profile.gs: an afternoon profile at the Viking Lander 1 site.

  
  
  

  
  

    

  Figure 2: Sample output from profile.gs: a profile in the Southern polar night.

  
  
  
• section_lat.gs This script is similar to profile.gs, except that a range of latitudes can be given to produce a sigma-latitude section through the database.

  The longitude, first and last latitude and local True Solar Time are supplied as arguments. If the latitude range given is too small, it is expanded to include several database gridpoints; otherwise profile.gs could be used. The longitude should be given in degrees East and the time in hours (00..24). This script must be edited to select a dust scenario and season and to change the title accordingly.

  GrADS should be run in portrait mode.

• zonal_tuv.gs Plots the zonal mean and diurnal time mean temperature, zonal and meridional velocity for one season in a $\ln (\sigma)$-latitude coordinate system. GrADS should be run in portrait mode for this script.

  Sample output from this script is given in Figures 3 and 4, which also allow a comparison of data from the ``standard'' and ``bright'' dust properties datasets.

  
  

    

  Figure 3: Sample output from zonal_tuv.gs, using data from the ``standard dust'' dataset, data_s.

  
  
  

  
  

    

  Figure 4: Sample output from zonal_tuv.gs, using data from the ``bright dust'' dataset, data_b.

  
  
  
• zonal_sdtuv.gs Plots the zonal mean standard deviation of temperature, zonal and meridional velocity for one season. GrADS should be run in portrait mode for this script.

GrADS is a popular and powerful public domain software package for displaying meteorological variables. You are referred to the manual for a description of all the features available. Also note that as GrADS is widely used there may be scripts already written to perform the particular task you are undertaking. It may be worth a search of the Internet and related newsgroups for GrADS scripts. The World Wide Web page at http://grads.iges.org/grads is a good starting place.